Processing load during speech perception in challenging listening conditions – insights from pupillometry.

Dr. Adriana E. Zekveld (Dept. of ENT / Audiology and the EMGO+ Institute for Health and Care Research, VU University medical center, Amsterdam, The Netherlands; Department of Behavioural Sciences and Learning, Linköping University, Sweden; Linnaeus Centre HEAD, The Swedish Institute for Disability Research, Linköping and Örebro Universities, Sweden)

6 novembre 2012, 16h, Uni-Mail M5193

Background:
Pupillometry provides an objective method for measuring cognitive load (listening effort) during speech perception in adverse listening conditions. We assessed the influence of individual factors like age, hearing loss, and cognitive ability and external factors like stimulus type, stimulus modality, intelligibility level, and masker type on the pupil response.
Methods:
Combining and comparing the results of 4 studies allowed us to assess the influence of speech intelligibility level and masker type on cognitive processing load during listening and how this relates to cognitive ability. Five groups of young (ntot = 123), two groups of middle-aged normal-hearing subjects (ntot = 62) and one group of middle-aged hearing-impaired subjects (ntot = 38) participated. Pupil dilation was recorded during speech perception in noise and during the text reception threshold (TRT) test. Sentences and words were presented over a wide range of sentence intelligibility levels (1% - 99% correct), masked by different masker types (steady-state, fluctuating, single-talker) and in quiet.
Results:
The results show consistently larger pupil responses in less intelligible or more complex listening conditions. The pupil response increases with decreasing speech intelligibility level. However, when speech perception is very difficult and subjects start to give up trying to perceive the speech, the pupil response decreases relative to less difficult conditions. This inverse-U shaped function of the pupil response across intelligibility levels supports the validity of the pupil response as measure of cognitive processing load. Middle-aged and hearing-impaired adults show a smaller decrease in pupil response with increasing intelligibility. Independent of intelligibility level, masker type affects cognitive processing load. Processing load is highest for speech masked by an interfering speaker. The TRT test is associated with cognitive load during listening, especially in relatively difficult listening conditions. The data suggest that an individual’s cognitive abilities modulate the relation between speech intelligibility and processing load. The pupil response furthermore reflects processing load differences caused by different stimulus types (word versus sentence perception), task characteristics (detection versus identification) and stimulus modality.
Conclusions:
The pupil response and speech perception performance data reflect different aspects of processing load during listening. The impact of several factors and individual differences that affect the complexity of the listening task and cognitive processing load are quantifiable by assessment of the pupil response during listening.